Why we want to sequence your ass

I’ve been rounding in the hospital on about 30 very sick patients every day for 2 weeks as attending physician on the Leukemia and Stem Cell Transplant Service. It’s one intense experience. I enjoy the intellectual challenge of supervising a team of doctors and nurses dealing with hundreds of extremely complex medical decisions. They only give me the hard ones. Emotionally, the experience is even more challenging. Patients on our service have fatal, nasty diseases, but stem cell transplant medicine brings them here hoping for a cure. Facing these patients and their families when things don’t go well, when the disease relapses, when complications set in is the hardest part of this job. As the attending doc, the buck stops with me. The staff, patients and families look to me for answers to impossible questions.

My colleagues and I have been using new sequencing technology to try and cure cancer, the same tech has allowed scientists to study our microbiomes: the bacteria, fungi and viruses that live all over us and in us. It’s super-cool science, because microbiologists have traditionally relied on culture to identify germs, but “metagenomics” allows characterization of our beasties straight away and therefore might be able to give a more detailed and accurate picture of the life on our bodies. I’m jaded when it comes to the claims of scientists regarding “translation” to clinical care in the near future. But in this case, we need this technology in the clinic right away.

Our patients have immune systems that have been devastated by bone marrow failure syndromes, leukemia and our treatments. These folks have severe immunosuppression and are susceptible to opportunistic bacterial, fungal and viral infections like nobody’s business.

I tell the fellows on bone marrow transplant the differential diagnosis of ANY symptom is “sepsis, sepsis, sepsis and sepsis.”


Clostridium difficile, a nasty bacteria that causes diarrhea and colitis in patients whose normal intestinal flora have been disrupted by antibiotics and immune suppression.

I approached Dr. George Weinstock at a meeting and told him that if he wanted to see huge shifts in human microbiota, he should check out our hematopoietic stem cell transplant patients. Today, we treat our patients largely empirically–often, we don’t know what infections we’re treating, we use “gorillacilin,” broad-spectrum antibiotics that target germs like a shotgun.

George was intrigued, and we gathered additional colleagues and we are starting to collaborate. Dr. Mark Schroeder another bone marrow transplant physician-researcher at Wash U is heading up the clinical trial. Our plan at this stage is to characterize the microbiomes of patients undergoing stem cell transplants and get a feel for what the technology can detect using real clinical specimens.

There is a young man with aplastic anemia right now very sick with infection by an unknown organism(s). We are giving him broad-spectrum antibiotics, but his immune system is destroyed and lack of white blood cells has made him a target for all variety of opportunistic organisms. We are trying to keep the man alive until we can get him his stem cell transplant–he needs a new immune system. Why can’t we use state-of-the-art DNA sequencing technology to tailor his antibiotic regimen today? Target bad germs early and specifically? The situation is desperate. Why is this technology not here where it needs to be?

Listening to what cancer genome sequencing is really telling us

A fascinating story came out of two papers published in Science Translational Medicine recently, (here and here) but didn’t get as much attention as they deserved IMHO because they don’t fit the “race for the cure” cancer research meme. Both groups used high-throughput genome sequencing to detect the “mutation signature” of a specific plant carcinogen, aristolochic acid (AA). This work joins a small list of other types of environmental exposures, e.g. sunlight and tobacco that have gene signatures that can be detected by DNA sequencing.

AA first came to prominence as a potent nephrotoxin and carcinogen through its inadvertent inclusion in an herbal remedy administered by a weight loss clinic in Belgium in the early 1990’s. Soon thereafter, Balkan endemic nephropathy, a condition associated with urothelial carcinoma of the upper urinary tract (UTUC) in rural populations along the River Danube, was found to be caused by accidental contamination of harvested wheat with the weed Aristolochia, which contains AA. —   Genomics Traces Carcinogen Fingerprints, William Lee and Marc Ladanyi

ImageAristolochia elegans

Devra Davis in her book “The Secret History of the War on Cancer”  tells the story of how US research enthralled by the pharmaceutical industry has almost systematically refused to look at the environmental causes of cancer. As a doctor specializing in bone marrow transplant, I must admit I always thought prevention was for weenies.  We deal with diseases (leukemia, lymphoma, myeloma) that have generally not been considered preventable. Prevention anyway is difficult– our tools (mammograms, PSA tests) are clumsy and inaccurate. On the other hand, understanding how advanced cancers work and finding “the cure” is an easy-to-understand concept that has won the vast majority of research funding.  And the “search for the cure” has dominated the narrative of science communication.

Many papers have used cancer genome sequencing in an effort to find specific mutations that drive cancer development.  The “race for the cure” meme dictates that the reason we care is so that we can find drugs to block these specific cancer driver mutations.

Studying cancer genome sequences several years ago, I wrote that ” Our genomes…have not received the drug company memos.” I meant that our intense desire to use sequencing experiments to find a “magic bullet,” was preventing us from learning important lessons. Could we stop for a minute and listen to what the data were telling us? To my mind, data were telling us, in the words of Clifton Leaf, we must “com[e] to terms with the fact that cancer is a disease of progressive, unyielding, mind-boggling heterogeneity. There is no magic bullet for that.”

These new papers are interesting because rather than looking for specific mutations, they looked at the DNA in these tumors broadly as a kind of “radio antenna” to detect a general type of mutation. They found that UTUC tumors were filled with not one mutation, but a class of mutation, A>T transversions on the non-transcribed strand.  The tumors that arose because of exposure to AA did not necessarily have one particular gene mutated, but they were filled with thousands of these specific types of mutations, providing a signature of exposure to this carcinogen.

But despite a growing list of new treatments, our progress remains incremental at best.  Cancer is a heterogeneous set of diseases that are tremendously resistant to cures once they become advanced. It breaks my heart to see the suffering and death caused by cancer and to feel that we might be misdirecting our efforts and

And the latest data from genome sequencing, if we are willing to listen, is telling us a story of cancer as a disease of environmental exposures.  And perhaps instead of looking for magic bullets, we should be looking at the processed foods and chemicals we surround ourselves with.

Addendum: A third paper just released surveys many cancer types and finds additional mutational signatures possibly related to carcinogens: Signatures of mutational processes in human cancer

Healthcare reform will kick your translational research in the teeth. In a good way.


Multiple myeloma patients usually respond well to initial treatment, and researchers are trying to figure out what new drugs to give when people relapse. Tremendous resources are applied here, late in the disease stage, and the sky is the limit. No expense is spared to cure people dying of cancer, after all.

There is a pre-malignant (MGUS) stage before patients get multiple myeloma. We used to think that the diagnosis was a lightning bolt that came out of the blue and there is no way to catch it until it happens. I have often been asked, “could my doctor have caught this earlier?” and I’ve often replied that “no, there was no way this could have been anticipated.” But, now I know that is no longer true. We could have detected it earlier. A year ago we could have detected it. A blood test that looks for abnormal antibodies – first thing that turns positive in the pre-malignant MGUS stage. It can be positive 8 years before they get myeloma.

Science is moving forward.  We know that every myeloma patient gets the pre-malignant phase first and we would see this if we looked. Is there anything we could have done before the development of full-blown myeloma? The answer now is “no” because not much research has been done regarding this early disease stage. The reason there is nothing we could do is because no one is fixated on this question. They are fixated on the cure.

The healthcare exchanges will be organized into tiers of coverage, platinum, gold, silver and bronze and I bet that these strata will start to influence medical research as well. Much of what we have been doing in the lab, honestly, is research into “platinum interventions–we assume money is no object. But many (most?) people will have silver and bronze coverage, where the cost versus benefit will very much be an object. What medical innovation would I pay for if it came out of my pocket? What should the plan cover?

That’s the point. Is there something we can do earlier that is less high tech but may be just as powerful.

Prevention is a bronze intervention, not platinum. Prevention is not a high tech laser beam or nano-technology intervention. Maybe it’s two aspirin a day. But, there’s not a lot of money in aspirin and so it’s difficult to do research. Clinical research is unbelievably expensive. For the most part, clinical trials are financed by people expecting a return on investment.

The “translational research’ community focused on high technology and new drug targets has been by and large insulated from shifting economic realities. Occasionally, I will hear someone ask half-heartedly whether a new approach would be cost-effective, but let’s face it, we are largely unconcerned with the expense side of the equation. Maybe that’s how it should be–unfettered imagination for disease research. And yet…

This is my motivation for working on cheap prevention strategies for myeloma. Our colleagues found recently that nearly all myeloma patients go through the pre-malignant MGUS stage first, and yet physicians have no treatments for MGUS patients to prevent them from progressing to multiple myeloma. It’s a numbers game…most MGUS patients will be fine. I want to know if it is possible to develop strategies to prevent myeloma from happening in the first place. How effective would prevention strategies have to be for them to work? Could we use modern genomics to focus our prevention efforts? This is why we are now working on a mathematical model for myeloma development at the population level. Healthcare reform is flipping the incentive structure in medicine. We are moving into a world where keeping people healthy will be the way to make money in healthcare. Soon, the shifting incentives will start to impact translational research. Stay tuned.